Traffic Injury Prevention最新文献

{"title":"A comparison between bridge-related and non-bridge crashes: severity, vehicle involvement, and motorists’ characteristics","authors":"Tung T. Tran ,&nbsp;Long T. Truong ,&nbsp;Athanasios Theofilatos","doi":"10.1080/15389588.2024.2400234","DOIUrl":"10.1080/15389588.2024.2400234","url":null,"abstract":"<div><h3>Objectives</h3><div>Bridges, though representing a small portion of transport networks by length, are still prone to traffic crashes. Despite extensive research on bridge-related crashes globally, there’s a scarcity of studies exploring differences between bridge- and non-bridge-related crashes. Thus, this paper attempts to add to the current knowledge by comparing bridge-related crashes and non-bridge crashes using a case study in Victoria, Australia.</div></div><div><h3>Methods</h3><div>By utilizing crash data in Victoria between January 2006 and April 2023; the Chi-squared test is conducted to examine differences between crash types. Partial proportion odds models are then employed to establish if bridge-related crashes tend to be more severe.</div></div><div><h3>Results</h3><div>In general, the analysis reveals that bridge-related crashes tend to result in more severe injury outcomes compared to non-bridge crashes. Furthermore, there is evidence that bridge-related crashes are more likely to be single-vehicle, to involve a heavy vehicle as well as young/male motorists, and to occur on high-speed roads, regional areas and at nighttime than non-bridge crashes.</div></div><div><h3>Conclusions</h3><div>A targeted program aiming at enhancing traffic safety at bridges is recommended due to the higher severity of bridge-related crashes. Considering the presence of bridges on the road network, and particularly in regional areas where many are in poorer conditions, it is essential to consider retrofitting as well as low-cost solutions in order to complement motorist education and awareness programs. Specific groups with higher risks of being involved in bridge-related crashes, such as young male motorists, should be targeted in these programs.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 191-197"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142481008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Road traffic crash data in Southwestern Nigeria: a qualitative analysis of sources, contents and management methods","authors":"Jude Ubaka Odinfono ,&nbsp;Moses Olaniran Olawole","doi":"10.1080/15389588.2024.2398670","DOIUrl":"10.1080/15389588.2024.2398670","url":null,"abstract":"<div><h3>Objective</h3><div>This study qualitatively examined road traffic crash (RTC) data collection and management in southwest Nigeria, with the goal of addressing the tenets of the UN Decade of Action Plan on Road Safety’s call for an effective data system.</div></div><div><h3>Methods</h3><div>Data on RTCs data collection in the study area was obtained through key informant interviews with five hospital accident emergency unit directors, three Federal Road Safety Commission (FRSC) Sector commanders, and three Nigeria Police Force (NPF) commissioners. Datasets were transcribed, categorized, and interpreted using content analysis, descriptive statistics, and WHO recommended minimal crash data element requirements.</div></div><div><h3>Results</h3><div>The study found disparate systems, a lack of synergy, and discrepancy in the various data sources in the country when compared with WHO recommendations on road traffic data collection systems, with 55% of agencies using non-standardized paper forms. The study also reveals that the minimum crash data elements recommended by WHO are not consistently captured in the NPF, FRSC, and hospital traffic crash data sources.</div></div><div><h3>Conclusions</h3><div>The study suggested the use of an upgraded National Road Traffic Crash Data Management System (NRTCDMS) Data Template to unify data gathering and linkage issues, but recommends a digital version of the template.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 236-242"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142481018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of traffic accident risk based on vehicle trajectory data","authors":"Hao Li ,&nbsp;Lina Yu","doi":"10.1080/15389588.2024.2402936","DOIUrl":"10.1080/15389588.2024.2402936","url":null,"abstract":"<div><h3>Objective</h3><div>The objective of this study is to conduct precise risk prediction of traffic accidents using vehicle trajectory data.</div></div><div><h3>Methods</h3><div>For urban road and highway scenarios, a scheme was developed to gather vehicle kinematic data and driving operation records from an in-vehicle device. The raw trajectory samples of over 3000 vehicles were processed through cleaning, filtering, interpolation, and normalization for preprocessing. Three deep learning frameworks based on RNN, CNN, and LSTM were compared. An end-to-end LSTM accident risk prediction model was constructed, and the model was trained using the cross-entropy loss function with Adam optimizer.</div></div><div><h3>Results</h3><div>The LSTM model is capable of directly extracting accident-related hazardous state features from low-quality raw trajectory data, thereby enabling the prediction of accident probability with fine-grained time resolution. In tests conducted under complex traffic scenarios, the model successfully identifies high-risk driving behaviors in high-speed road sections and intersections with a prediction accuracy of 0.89, demonstrating strong generalization performance.</div></div><div><h3>Conclusions</h3><div>The LSTM accident risk prediction model, based on vehicle trajectory, developed in this study, is capable of intelligently extracting dangerous driving features. It can accurately warn about the risk of traffic accidents and provides a novel approach to enhancing road safety.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 164-171"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142683577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Investigation of a surrogate measure-based safety index for predicting injury crashes at signalized intersections","authors":"Maryam Hasanpour ,&nbsp;Bhagwant Persaud ,&nbsp;Robert Mansell ,&nbsp;Craig Milligan","doi":"10.1080/15389588.2024.2397652","DOIUrl":"10.1080/15389588.2024.2397652","url":null,"abstract":"<div><h3>Objectives</h3><div>The paper develops a machine learning-based safety index for classifying traffic conflicts that can be used to estimate the frequency of signalized intersection crashes, with a focus on the more severe ones that result in fatal and severe injury. The number of conflicts in different severity levels categorized by the safety index is used as an explanatory variable for developing statistical models for pro-actively estimating crashes.</div></div><div><h3>Methods</h3><div>Video-derived conflicts in different severity levels between left-turning vehicles and opposing through vehicles, a well-recognized severe injury crash typology at signalized intersections, were identified by jointly integrating the indicators of frequency and severity, using an autoencoder neural network integration method to develop anomaly scores. Regression models were then developed to relate crashes at the same intersections to the classified conflicts based on the value of their safety indexes. Cumulative Residual plots were investigated. Finally, equations defining the boundary between consecutive anomaly score levels were developed to facilitate application in practice.</div></div><div><h3>Results</h3><div>Regression models for total and fatal plus severe (FSI) crashes utilizing classified extreme conflicts based on anomaly scores were found to outperform the models using total conflicts. The improvement is more pronounced for FSI crashes. The results also suggest that the machine learning integration method can efficiently classify conflicts accurately according to crash severity levels since the higher anomaly score is associated with a higher crash severity level (i.e., FSI).</div></div><div><h3>Conclusions</h3><div>The proposed framework represents a methodological advancement in traffic conflict-based estimation of crashes using a machine learning model to classify conflicts by their anomaly scores. For jurisdictions without the resources to develop such a model to classify conflicts for their own datasets, the simple equations defining the boundary between consecutive anomaly score levels could be used as an approximation.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 172-181"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142332372","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"High-risk powered two-wheelers scenarios generation for autonomous vehicle testing using WGAN","authors":"Xiaolong Luo ,&nbsp;Zhiyuan Wei ,&nbsp;Guoqing Zhang ,&nbsp;Helai Huang ,&nbsp;Rui Zhou","doi":"10.1080/15389588.2024.2399305","DOIUrl":"10.1080/15389588.2024.2399305","url":null,"abstract":"<div><h3>Objective</h3><div>Autonomous vehicles (AVs) have the potential to revolutionize the future of mobility by significantly improving traffic safety. This study presents a novel method for validating the safety performance of AVs in high-risk scenarios involving powered 2-wheelers (PTWs). By generating high-risk scenarios using in-depth crash data, this study is devoted to addressing the challenge of public road scenarios in testing, which often lack the necessary complexity and risk to effectively evaluate the capabilities of AVs in high-risk situations.</div></div><div><h3>Method</h3><div>Our approach employs a Wasserstein generative adversarial network (WGAN) to generate high-risk scenes, particularly focusing on PTW scenarios. By extracting 314 car-to-PTW crashes from the China In-depth Mobility Safety Study–Traffic Accident database, we simulate outcomes using PC-Crash software. The data are divided into scenes at 0.1-s intervals, with WGAN generating numerous high-risk scenes. By using a cumulative distribution function (CDF), we sampled and analyzed the vehicle’s dynamic information to generate complete scenarios applicable to the test. The validation process involves using the SVL Simulator and the Baidu Apollo joint simulation platform to evaluate the AV’s driving behavior and interactions with PTWs.</div></div><div><h3>Results</h3><div>This study evaluates model generation results by comparing distributions using Wasserstein distance as an indicator. The generator converges after approximately 200 epochs, with the iterator converging quickly. Subsequently, 10,000 new scenes are then generated. The distribution of several key parameters in the generated scenes can be found to approximate that of the original scenes. After sampling, the usability of generated scenarios is 64.76%. Virtual simulations confirm the effectiveness of the scenario generation method, with a generated scenario crash rate of 16.50% closely reflecting the original rate of 15.0%, showcasing the method’s capacity to produce realistic and hazardous scenarios.</div></div><div><h3>Conclusions</h3><div>The experimental results suggest that these scenarios exhibit a level of risk similar to the original crashes and are effective for testing AVs. Consequently, the generated scenarios enhance the diversity of the scenario library and accelerate the overall testing process of AVs.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 243-251"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142481012","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Supporting driver performance in closely spaced tunnel-interchange structure: Traffic control devices and effects on driving behavior","authors":"Yunjie Ju ,&nbsp;Feng Chen ,&nbsp;Jia Li ,&nbsp;Xiaohua Zhao ,&nbsp;Wenhui Dong","doi":"10.1080/15389588.2024.2401495","DOIUrl":"10.1080/15389588.2024.2401495","url":null,"abstract":"<div><h3>Objectives</h3><div>With the rapid development of expressways in the mountainous regions of southwestern China, closely spaced tunnel-interchange structures have inevitably emerged due to topographical constraints and environmental limitations. Given the unfavorable road geometry and rapid cross-section transitions, drivers face significant safety concerns. This study aims to investigate drivers’ safety performance at closely spaced tunnel-interchange sections and determine how safety risks can be mitigated through improved traffic control devices design.</div></div><div><h3>Methods</h3><div>Thirty-nine participants conducted an experimental study in a fixed-base simulator. The test scenario was modeled on the Xingyan Freeway-S3801 and accurately reproduced in the simulator. For each safety performance metric, the driving simulator experiments yielded a dataset with 780 observations. To address the idiosyncratic variation due to individual driver differences, a series of linear mixed effects models (LMM) were developed to analyze drivers’ behavior responses.</div></div><div><h3>Results</h3><div>In closely spaced tunnel-interchange sections, a general impairment of both longitudinal and lateral performance was observed. This study identified potential critical impact variables in traffic control device systems. According to the LMM results: (a) Removing the 0.5 km interchange ramp exit advance guide sign located in the tunnel exit area reduces dangerous behavior in the corresponding impact area. (b) Replacing the 0.5 km interchange ramp exit advance guide sign with arrow pavement markers as an information source supports improved driver performance, promoting driver safety. (c) Adding tunnel exit distance signs within tunnels is recommended to enhance situation awareness for drivers.</div></div><div><h3>Conclusions</h3><div>This study addresses the scientific issues related to traffic control devices setup for closely spaced tunnel-interchange sections, focusing on identifying potential critical impact variables. The findings provide guidance on the design of traffic control devices for such sections and support revisions to national engineering standards.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 207-214"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142367362","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting pedestrian-vehicle interaction severity at unsignalized intersections","authors":"Kaliprasana Muduli ,&nbsp;Indrajit Ghosh","doi":"10.1080/15389588.2024.2404713","DOIUrl":"10.1080/15389588.2024.2404713","url":null,"abstract":"<div><h3>Objectives</h3><div>This study aims to develop and validate a novel deep-learning model that predicts the severity of pedestrian-vehicle interactions at unsignalized intersections, distinctively integrating Transformer-based models with Multilayer Perceptrons (MLP). This approach leverages advanced feature analysis capabilities, offering a more direct and interpretable method than traditional models.</div></div><div><h3>Methods</h3><div>High-resolution optical cameras recorded detailed pedestrian and vehicle movements at study sites, with data processed to extract trajectories and convert them into real-world coordinates <em>via</em> precise georeferencing. Trained observers categorized interactions into safe passage, critical event, and conflict based on movement patterns, speeds, and accelerations. Fleiss Kappa statistic measured inter-rater agreement to ensure evaluator consistency. This study introduces a novel deep-learning model combining Transformer-based time series data capabilities with the classification strengths of a Multilayer Perceptron (MLP). Unlike traditional models, this approach focuses on feature analysis for greater interpretability. The model, trained on dynamic input variables from trajectory data, employs attention mechanisms to evaluate the significance of each input variable, offering deeper insights into factors influencing interaction severity.</div></div><div><h3>Results</h3><div>The model demonstrated high performance across different severity categories: safe interactions achieved a precision of 0.78, recall of 0.91, and F1-score of 0.84. In more severe categories like critical events and conflicts, precision and recall were even higher. Overall accuracy stood at 0.87, with both macro and weighted averages for precision, recall, and F1-score also at 0.87. The variable importance analysis, using attention scores from the proposed transformer model, identified ‘Vehicle Speed’ as the most significant input variable positively influencing severity. Conversely, ‘Approaching Angle’ and ‘Vehicle Distance from Conflict Point’ negatively impacted severity. Other significant factors included ‘Type of Vehicle’, ‘Pedestrian Speed’, and ‘Pedestrian Yaw Rate’, highlighting the complex interplay of behavioral and environmental factors in pedestrian-vehicle interactions.</div></div><div><h3>Conclusions</h3><div>This study introduces a deep-learning model that effectively predicts the severity of pedestrian-vehicle interactions at crosswalks, utilizing a Transformer-MLP hybrid architecture with high precision and recall across severity categories. Key factors influencing severity were identified, paving the way for further enhancements in real-time analysis and broader safety assessments in urban settings.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 252-261"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142481015","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Identifying the reciprocal causation between hit-and-run behavior and crash injury severity","authors":"Guopeng Zhang ,&nbsp;Qianwei Xuan ,&nbsp;Ying Cai ,&nbsp;Xianghong Hu ,&nbsp;Nianyi Hu ,&nbsp;Xinguo Jiang ,&nbsp;Xinkun Yao","doi":"10.1080/15389588.2024.2402464","DOIUrl":"10.1080/15389588.2024.2402464","url":null,"abstract":"<div><h3>Objective</h3><div>Hit-and-run behavior is believed to exacerbate the injury severity of traffic crashes due to the delayed emergency response for the victims. However, several previous studies indicated the opposite finding that hit-and-run crashes were associated with less severe injuries. The relevant studies mainly identified the statistical associations between hit-and-run behavior and injury severity without revealing causation between them. To this end, the study aims to explore the reciprocal causation between the two variables.</div></div><div><h3>Method</h3><div>The two-stage probit model with endogenous regressors is employed to identify the reciprocal causation between hit-and-run behavior and crash injury severity for single- and two-vehicle crashes, respectively, with the use of crash data extracted from the Crash Report Sampling System and Fatality Analysis Reporting System (2016-2019).</div></div><div><h3>Results</h3><div>The results indicate that 1) for both single- and two-vehicle crashes, the fleeing behavior can significantly increase the injury severity of the victims in the crashes while the severe injury of the victims has a negative impact on the propensity of such behavior, 2) the propensity of hit-and-run behavior is influenced by various instrumental variables such as driver age, gender, alcohol involvement, weekday, area type, and light condition, and 3) crash injury severity is significantly related to the victim age, gender, and vehicle damage.</div></div><div><h3>Conclusions</h3><div>There is a reciprocal causation between hit-and-run behavior and injury severity in traffic crashes. The analytical results can provide a reasonable explanation for the counterintuitive finding on hit-an-run crashes and help mitigate the injury severity.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 156-163"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142367360","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Acceptability of virtual restraint fitting to extend the reach of child restraint fitting services: A pilot randomized controlled trial","authors":"Nipuna Cooray ,&nbsp;Catherine Ho ,&nbsp;Wennie Dai ,&nbsp;Rebecca Szabo ,&nbsp;Kathy Tailor ,&nbsp;Miranda Maling ,&nbsp;Jason Chambers ,&nbsp;Sjaan Koppel ,&nbsp;Lynne Bilston ,&nbsp;Lisa Keay ,&nbsp;David Schwebel ,&nbsp;Julie Brown","doi":"10.1080/15389588.2024.2394966","DOIUrl":"10.1080/15389588.2024.2394966","url":null,"abstract":"<div><h3>Objective</h3><div>Incorrect use of child restraints is a long-standing issue, limiting the protection offered by child restraints in the event of a crash. Child restraint fitting services are a measure to reduce incorrect use but have limited reach and availability to underserved populations. Virtual child restraint fitting services have the potential to increase the reach and availability, but as with any digital intervention, need to be acceptable to users to be effective. The acceptability of such interventions has not been studied before.</div></div><div><h3>Methods</h3><div>Using a three-arm randomized controlled trial, this study evaluated the acceptability of: (1) a video with child restraint fitting advice (Control), (2) a traditional in-person child restraint fitting service (In-person), and (3) a virtual child restraint fitting service (Virtual). Additionally, the effectiveness in reducing incorrect use was evaluated.</div></div><div><h3>Results</h3><div>There was a significantly higher level of overall acceptability for the in-person service, and significantly fewer errors in child restraint use in this group compared to the control. There were no significant differences in overall acceptability or errors between the virtual service and the control. However in-depth analysis of the constructs of acceptability demonstrated participants in the in-person and virtual service groups held similar views on four of the seven constructs including the usefulness of the services and the impact of the service on comprehension of key information for correct restraint use. Areas where the views differed between these groups included perceived burden, appropriateness, and opportunity costs. Qualitative feedback suggested these negative perceptions on the virtual service may be remediated with some improvements to the technology.</div></div><div><h3>Conclusions</h3><div>Overall, child restraint fitting services provided virtually show promise as an alternative to in-person but attention to how services are provided <em>via</em> this technology, together with technology improvement, might be needed to fully realize its potential.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 146-155"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142803464","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An integrated framework for driving risk evaluation that combines lane-changing detection and an attention-based prediction model","authors":"Zhongxiang Feng ,&nbsp;Xinyi Wei ,&nbsp;Yu Bi ,&nbsp;Dianchen Zhu ,&nbsp;Zhipeng Huang","doi":"10.1080/15389588.2024.2399301","DOIUrl":"10.1080/15389588.2024.2399301","url":null,"abstract":"<div><h3>Objective</h3><div>In recent years, the increase in traffic accidents has emerged as a significant social issue that poses a serious threat to public safety. The objective of this study is to predict risky driving scenarios to improve road safety.</div></div><div><h3>Methods</h3><div>On the basis of data collected from naturalistic driving real-vehicle experiments, a comprehensive framework for identifying and analyzing risky driving scenarios, which combines an integrated lane-changing detection model and an attention-based long short-term memory (LSTM) prediction model, is proposed. The performance of the 4 machine learning methods on the CULane data set is compared in terms of model running time and running speed as evaluation metrics, and the ultrafast network with the best performance is selected as the method for lane line detection. We compared the performance of LSTM and attention-based LSTM on the basis of the prediction accuracy, recall, precision, and F1 value and selected the better model (attention-based LSTM) for risky scenario prediction. Furthermore, Shapley additive explanation analysis (SHAP) is used to understand and interpret the prediction results of the model.</div></div><div><h3>Results</h3><div>In terms of algorithm efficiency, the running time of the ultrafast lane detection network only requires 4.1 ms, and the average detection speed reaches 131 fps. For prediction performance, the accuracy rate of attention-based LSTM reaches 96%, the precision rate is 98%, the recall rate is 96%, and the F1 value is 97%.</div></div><div><h3>Conclusions</h3><div>The improved attention-based LSTM model is significantly better than the LSTM model in terms of convergence speed and prediction accuracy and can accurately identify risky scenarios that occur during driving. The importance of factors varies by risky scenario. The characteristics of the yaw rate, speed stability, vehicle speed, acceleration, and lane change significantly influence the driving risk, among which lane change has the greatest impact. This study can provide real-time risky scenario prediction, warnings, and scientific decision guidance for drivers.</div></div>","PeriodicalId":54422,"journal":{"name":"Traffic Injury Prevention","volume":"26 2","pages":"Pages 198-206"},"PeriodicalIF":1.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142367357","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}